EEVblog #888 - Mailbag

More Mailbag!
Australia's first mobile phone call:

SPOILERS:
Electronics Dummy Load and Dave discusses potential technical issues with paralelling MOSFET's
Atheist Book: Christian Mythology for Kids
Metal Rectifier, / Selenium Rectifier.
Hornby Triang model train P5A Power Controller.
Mitsubishi MT-398FOR6A Analog Mobile Phone
Swiss Tools Screwdrivers

Support the EEVblog through Patreon!

EEVblog Amazon Store (Dave gets a cut):

[kaizen]

The critique of that initial PCB was very very interesting Dave. Shame we can't get a full time series/segment.

Excellent video as always.

[djs]

"That's the one with all the shafts. Jeez, they're long too!" ~ Dave Jones 2016

[Graham_Langley]

Re. the dummy load: The mod I'd recommend is to split the circuit into two (or more) transconductance amps (that's the LM358, MOSFET and Rsense block) run in parallel. That's what I did when I first came up with the analogue version of this cct. back in the early '80s, although I used TIP121 Darlingtons as that's what the firm had in its parts inventory.

I've used the cct many times over the years, mainly in test equipment. Neatest one was a self-contained box that ran off the supply it was loading. With careful layout (on Veroboard) the load current was held within 1mA over a 6-30V / 10mA -5A range and it still gets regularly used over 30 years after it was built.

[vladtepesch]

hahaha - your facial expression on the first look on book - great :-D

[stranger]

35:04 For those interested, Dave already did a teardown of that Geiger counter in EEVblog #776.

[ScottHenion]

Those selenium rectifier had a distinct smell when they overheated. Something us old timers will remember.

[xlfisher]

Dave, Love the EEVblog. Not any good with electronics myself but I love your channel as I find it fascinating. On a side note a lot of European names beginning with J are pronounced as Y.

[MatthewHarrold]

My dad was a "sparky", my eldest son is doing Engineering/Mechatronics, I'm trying to keep up via this channel. Your voice and enthusiasm reminds me of Shirley Strachan (Skyhooks). I'm learning heaps. Cheers.

[KennethScharf]

Wow, that train power pack. I had one as a kid, but the variable voltage was controlled by a sliding tap on the single layer secondary of the transformer (sorta like a variac), but NOT on a toroid coil, the tranny was your typical I-E core. Another winding provided 22vac for the aux stuff. It also had a Se rectifier (FW bridge) and a DPDT switch to reverse direction.

Just a suggestion for the chap who designed the arduino protoshield. How about a footprint for the mega?

[followthetrawler]

Dave, thanks for the link to the Gtronics protoboard - I just ordered one, so their 'advertising' worked! Cant wait for it to arrive now!

[LiberatedMindcom]

Thanks for sharing! LOVE the look on your face before you realized it was a secular book. Priceless.

[ryanwilson]

dave genuino are genuine arduinos. arduino have changed how ever thing works

[userPrehistoricman]

As soon as I heard the description of that rectifier, I was thinking about selenium diodes. Photonicinduction destroyed some and was surprised at how little power it took.

[briannebeker]

MOSFET transistors are resistive devices (they do not act like diodes) and have a positive coefficient of temperature which prevents thermal runaway. Thus with two MOSFETs in parallel, the Rds of the MOSFETs rises with temperature so if one transistor has a lower Rds it will carry more current and thus it will heat up more than the other transistor, but both transistor still carry current. As long as the transistors are nominally fairly close in resistance they will thermally self balance. If one transistor has much better heat dissipation that the other it will conduct more current but because it will be more efficient thus it will do it with at a similar temperature. Using a common thermal path for both transistors does have the advantage that is one portion of the heatsink has better thermal flow then both transistors will benefit to some degree. Improving the thermal dissipation of the transistor with less efficient thermal flow will have a bigger impact on the total efficiency of the transistors than improving the thermal flow for the better thermally dissipated transistor. Power MOSFETs should generally not conduct more than 50-75% of their rated current because in addition to their current rating they need to be operated in their safe operating area. This also ensure that an imbalance of the current sharing should not cause the higher current transistor to exceed is safe operating area. If this imbalance is a result of a failure then the remaining transistor will run hotter but still within it operating specs.


BJT transistors lower the Vce with temperature which results in one transistor conducting all of the current because the Vce of the transistor is not reached. Putting BJTs on a common heat sink and using a emitter resistor allows the transistors to track better and ensures the Vce is met by both transistors.
Failing to do some could result in one transistor running cold while the other overheats due to thermal runaway.

It is however important to drive the gates of the MOSFETs properly due to the gate capacitance. Rapidly turning a MOSFET off that is carrying a large current requires discharging the gate capacitance and the pulling the gate to the same potential as the source. With large currents flowing ground bounce can leave the source pin at a lower potential than ground which can result in the transistor not fully turning off. Resistance and inductance in the driver circuits between the transistor driver and the source pins can cause the gate voltage to rise in reference to the source pin which can cause the gate to partially turn back on. The gate resistors thus can server 3 purposes.

The first thing the gate resistors do is slow down the rise and fall of the gate signal to lower high frequency turn-on and turn-off transients. The second purpose is to dissipate inducted inductive currents on the gate drive signal. The third purpose is that the gate resistor can prevent excess currents making their way back to the driver circuit if the MOSFET fails shorted to the drain. Using separate gate resistors balances the gate drive current so they both tend to turn on and turn off at the same speed and you can use higher resistance values for the resistor.

Thus a good gate driver for quick switching applications should have a large current sinking capability. Being about to sink 1 ampere or more means the driver can readily absorb gate bounce and can also survive current from a short from the drain to the gate. This also means the driver ground should be star grounded to the source pins. Of course the higher the switching frequency and the higher the currents switched the more important the gate driving becomes.

[PatrickPoet]

I'm so glad you sharpened your knife. It was breaking my heart.

[JerryWalker]

The un-smoothed controller was an intentional part of the 'design'. The average power drove the motor but the high peak voltage helped to cut through the dirt on the rails.

[CP]

I have a Triang-Hornby HO scale train set from about the same era... early to mid 1970's. The controller for it has a slide switch (forward/reverse) for direction and a sliding/rotary lever for speed control. How I loved that train set.... wonder if it still works?

[cipmars]

Wow, the full kit is only 70 euros! That's GREAT!!!

[bedad]

FSM! Whole new level of respect for you Dave.

[Phantomthecat]

The Triang power supply only has to run a DC motor, don't blow a gasket Dave! :). The 15 volt AC output was used for point motors - although you could use the 12 volt DC unregulated supply they worked better (at least the Hornby points did) on 15VAC. :)